Employment Relocation Model (ERM)

Objective

The Employment Relocation Model predicts the relocation of households within the region each simulation year.

Algorithm

The Employment Relocation Model is implemented as a cross-classification rate-based model, with a probability of moving by employment sector applied to each job, each simulation year. For example, if a job is in the retail sector, their probability of moving would be looked up by finding the retail sector entry in the annual_employment_relocation_rates table. Let's assume the rate in the table is .25. This means there is a 25% chance the job will move in any given year, and 75% chance they will not move in that year. The model uses Monte Carlo Sampling to determine the outcome. It works by drawing a random number (from the uniform distribution, between 0 and 1), and comparing that random draw to the probability of moving for each household. So with our example job's probability of 0.75 that they will stay, if we draw a random number with a value higher than 0.75, we will predict that the job will move in that year.

The outcome of the model is implemented as follows. If a job is determined to be a mover because the random draw is greater than (1 - their move probability), then they are moved out of their current location. In practical terms, their building_id, which identifies where they are located, is simply reset to a null value. They remain in the jobs table but do not have a location.

Configuration

The configuration of the ERM is summarized in the following table:

Element	Setting
Agent	Job
Dataset	Job
Model Structure	Cross-classification rate-based Model

Data

The following tables are used in the Employment Relocation Choice model:

Table Name	Brief Description
annual_job_relocation_rates	Annual relocation rates for jobs by sector
jobs	jobs

-- PaulWaddell - 09 Dec 2009